An agentic pipeline for contract, case, and matter analysis. Not a chatbot — a production system with per-matter memory, a verification gate, and inspectable reasoning every partner can audit.
Pillar 01Live in productionUK + global EnglishPer-matter isolation
What it does
— 01
Domain-specialised agents identify indemnification, limitation-of-liability, change-of-control, IP assignment, and jurisdictional clauses. Each flagged with a composite risk score against the firm's house rubric.
— 02
Automatically classifies incoming matter documents into practice-area taxonomies: commercial dispute, corporate transaction, employment, regulatory. Routes to the right senior reviewer.
— 03
Per-firm source-of-truth memory pulls relevant historic matters, house positions, and negotiated carve-outs — so the reasoning is grounded in the firm's own playbook, not a generic web corpus.
— 04
Draft clauses and position memos produced by the drafting agent, then audited by a separate verification agent against the firm's standards before human review. Confidence score attached to every output.
— 05
Matter-level isolation layer ensures lawyer-client privilege is preserved across agent interactions. No cross-matter bleed. Full audit trail for compliance.
— 06
Every decision carries a full trace: which agent produced it, which sources it consulted, which verification rules ran. Produced on demand for professional-conduct reviews or insurer queries.
Who it’s for
Mid-sized commercial law firms
The firms already feeling margin pressure on associate billing, looking for ways to handle more matters per fee-earner without compromising partner liability. Current alternatives: Harvey, Spellbook, Luminance. Blackflake's difference — inspectable agents you license, not a sealed enterprise SaaS billed per seat.
In-house legal teams at scale-ups
The fast-growing company that hires one GC and then drowns in NDAs, MSAs, employment agreements, and vendor contracts. Legal AI triages and risk-scores incoming contracts so the GC's time concentrates on the 10% that actually need judgment.
How it’s built
Not a single prompt or a chatbot. A composed pipeline of specialised agents, each with a defined role, a documented toolset, and a declared failure mode.
agent/ingest-matterParses the incoming contract, case bundle, or matter pack. Identifies matter type, parties, jurisdiction, and governing law. Produces a structured matter record.
agent/extract-clausesDomain-tuned extraction of load-bearing clauses — indemnity, limitation, IP, non-compete, change-of-control, force majeure, dispute resolution.
agent/score-riskAssigns a composite risk score per clause against the firm's house rubric and precedent corpus. Flags outliers requiring partner attention.
agent/retrieve-precedentQueries the per-matter source of truth — prior matters for this client, firm house positions, negotiated carve-outs. Reasoning is grounded in the firm's own playbook.
agent/qa-gate-v1Verification gate. Independent agent audits every draft and risk score against firm-specific rules and known failure modes. Rejects or flags anything below confidence threshold.
agent/compose-memoProduces the draft position memo or advice note, ready for partner review. Full reasoning trail attached.
Status
Status
In production. Deployed across live matters since early 2026.
Availability
Engagement by scoped pilot (6–12 weeks) or direct licence.
Jurisdictions
UK (England & Wales primary), EU English-language work.
Compliance
UK GDPR & EU GDPR compliant. Per-matter isolation. Full DPA on request.
Commercial model
Per-firm licence, not per-seat. Volume tiers by fee-earner count.
Engage
enterprise@blackflake.com with matter volume and practice-area focus.
How we engage
Every Legal AI deployment is a bespoke engagement. We don't sell seats or generic SaaS. We build the agents against your workflows, your matter taxonomy, your regulatory context — then we operate the system with you under a monthly retainer. You own the deployed build.
— 01
Scoped call to map your workflow, constraints, and success metrics. No charge. You leave with an architecture note and engagement estimate.
— 02
Fixed-price deployment on one narrow workflow. Real output on real work in weeks. Go / no-go before the full build.
— 03
Agent suite, workflows, sources of truth, QA gates, operator surface. Deployed in your tenant. Auditable.
— 04
Monthly retainer: new regulations ingested, agents tuned on live work, new workflows added. We develop the engine; your data, configurations, and outputs stay yours — always portable.